Skip to content

[Headers][X86] Allow AVX movddup/movsldup/movshdup intrinsics to be used in constexpr #152340

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Aug 7, 2025
Merged

[Headers][X86] Allow AVX movddup/movsldup/movshdup intrinsics to be used in constexpr #152340

merged 1 commit into from
Aug 7, 2025

Conversation

RKSimon
Copy link
Collaborator

@RKSimon RKSimon commented Aug 6, 2025

Matches SSE3 handling

@RKSimon RKSimon requested a review from phoebewang August 6, 2025 16:59
@llvmbot llvmbot added clang Clang issues not falling into any other category backend:X86 clang:headers Headers provided by Clang, e.g. for intrinsics labels Aug 6, 2025
@llvmbot
Copy link
Member

llvmbot commented Aug 6, 2025 •
edited
Loading

@llvm/pr-subscribers-clang

@llvm/pr-subscribers-backend-x86

Author: Simon Pilgrim (RKSimon)

Changes

Matches SSE3 handling

Full diff: https://github.com/llvm/llvm-project/pull/152340.diff

4 Files Affected:

  • (modified) clang/lib/Headers/avx512fintrin.h (+3-3)
  • (modified) clang/lib/Headers/avxintrin.h (+3-3)
  • (modified) clang/test/CodeGen/X86/avx-builtins.c (+3)
  • (modified) clang/test/CodeGen/X86/avx512f-builtins.c (+3)
diff --git a/clang/lib/Headers/avx512fintrin.h b/clang/lib/Headers/avx512fintrin.h
index 45e7eeb5327d0..b743c5b8424a5 100644
--- a/clang/lib/Headers/avx512fintrin.h
+++ b/clang/lib/Headers/avx512fintrin.h
@@ -5303,7 +5303,7 @@ _mm512_mask_store_epi64 (void *__P, __mmask8 __U, __m512i __A)
           (__mmask8) __U);
 }
 
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
+static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR
 _mm512_movedup_pd (__m512d __A)
 {
   return (__m512d)__builtin_shufflevector((__v8df)__A, (__v8df)__A,
@@ -8665,7 +8665,7 @@ _mm512_mask_testn_epi64_mask (__mmask8 __U, __m512i __A, __m512i __B)
                                        _mm512_setzero_si512());
 }
 
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
+static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR
 _mm512_movehdup_ps (__m512 __A)
 {
   return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
@@ -8688,7 +8688,7 @@ _mm512_maskz_movehdup_ps (__mmask16 __U, __m512 __A)
                                              (__v16sf)_mm512_setzero_ps());
 }
 
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
+static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR
 _mm512_moveldup_ps (__m512 __A)
 {
   return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
diff --git a/clang/lib/Headers/avxintrin.h b/clang/lib/Headers/avxintrin.h
index b9ca013c25c7a..0ccdaaa1f820b 100644
--- a/clang/lib/Headers/avxintrin.h
+++ b/clang/lib/Headers/avxintrin.h
@@ -2392,7 +2392,7 @@ _mm256_cvtss_f32(__m256 __a)
 ///    return value.
 /// \returns A 256-bit vector of [8 x float] containing the moved and duplicated
 ///    values.
-static __inline __m256 __DEFAULT_FN_ATTRS
+static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm256_movehdup_ps(__m256 __a)
 {
   return __builtin_shufflevector((__v8sf)__a, (__v8sf)__a, 1, 1, 3, 3, 5, 5, 7, 7);
@@ -2417,7 +2417,7 @@ _mm256_movehdup_ps(__m256 __a)
 ///    return value.
 /// \returns A 256-bit vector of [8 x float] containing the moved and duplicated
 ///    values.
-static __inline __m256 __DEFAULT_FN_ATTRS
+static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm256_moveldup_ps(__m256 __a)
 {
   return __builtin_shufflevector((__v8sf)__a, (__v8sf)__a, 0, 0, 2, 2, 4, 4, 6, 6);
@@ -2439,7 +2439,7 @@ _mm256_moveldup_ps(__m256 __a)
 ///    the return value.
 /// \returns A 256-bit vector of [4 x double] containing the moved and
 ///    duplicated values.
-static __inline __m256d __DEFAULT_FN_ATTRS
+static __inline __m256d __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm256_movedup_pd(__m256d __a)
 {
   return __builtin_shufflevector((__v4df)__a, (__v4df)__a, 0, 0, 2, 2);
diff --git a/clang/test/CodeGen/X86/avx-builtins.c b/clang/test/CodeGen/X86/avx-builtins.c
index a6e70aae420ea..539e6773998ba 100644
--- a/clang/test/CodeGen/X86/avx-builtins.c
+++ b/clang/test/CodeGen/X86/avx-builtins.c
@@ -1258,18 +1258,21 @@ __m256d test_mm256_movedup_pd(__m256d A) {
   // CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
   return _mm256_movedup_pd(A);
 }
+TEST_CONSTEXPR(match_m256d(_mm256_movedup_pd((__m256d){+7.0, -7.0, -42.0, +42.0}), +7.0, +7.0, -42.0, -42.0));
 
 __m256 test_mm256_movehdup_ps(__m256 A) {
   // CHECK-LABEL: test_mm256_movehdup_ps
   // CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
   return _mm256_movehdup_ps(A);
 }
+TEST_CONSTEXPR(match_m256(_mm256_movehdup_ps((__m256){+1.0f,-1.0f,+2.0f,+4.0f,+8.0f,-8.0f,-3.0f,+3.0f}), -1.0f, -1.0f, +4.0f, +4.0f, -8.0f, -8.0f, +3.0f, +3.0f));
 
 __m256 test_mm256_moveldup_ps(__m256 A) {
   // CHECK-LABEL: test_mm256_moveldup_ps
   // CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
   return _mm256_moveldup_ps(A);
 }
+TEST_CONSTEXPR(match_m256(_mm256_moveldup_ps((__m256){+1.0f,-1.0f,+2.0f,+4.0f,+8.0f,-8.0f,-3.0f,+3.0f}), +1.0f, +1.0f, +2.0f, +2.0f, +8.0f, +8.0f, -3.0f, -3.0f));
 
 int test_mm256_movemask_pd(__m256d A) {
   // CHECK-LABEL: test_mm256_movemask_pd
diff --git a/clang/test/CodeGen/X86/avx512f-builtins.c b/clang/test/CodeGen/X86/avx512f-builtins.c
index 84e700cfbd378..20d2662d0082e 100644
--- a/clang/test/CodeGen/X86/avx512f-builtins.c
+++ b/clang/test/CodeGen/X86/avx512f-builtins.c
@@ -4386,6 +4386,7 @@ __m512d test_mm512_movedup_pd(__m512d __A) {
   // CHECK: shufflevector <8 x double> %{{.*}}, <8 x double> %{{.*}}, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
   return _mm512_movedup_pd(__A);
 }
+TEST_CONSTEXPR(match_m512d(_mm512_movedup_pd((__m512d){-1.0, +2.0, +3.0, +4.0, -5.0, -6.0, +7.0, +8.0}), -1.0, -1.0, +3.0, +3.0, -5.0, -5.0, +7.0, +7.0));
 
 __m512d test_mm512_mask_movedup_pd(__m512d __W, __mmask8 __U, __m512d __A) {
   // CHECK-LABEL: test_mm512_mask_movedup_pd
@@ -8682,6 +8683,7 @@ __m512 test_mm512_movehdup_ps(__m512 __A) {
   // CHECK: shufflevector <16 x float> %{{.*}}, <16 x float> %{{.*}}, <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
   return _mm512_movehdup_ps(__A);
 }
+TEST_CONSTEXPR(match_m512(_mm512_movehdup_ps((__m512){+1.0f,-1.0f,+2.0f,-2.0f,+3.0f,-3.0f,+4.0f,-4.0f,+5.0f,-5.0f,+6.0f,-6.0f,+7.0f,-7.0f,+8.0f,-8.0f}), -1.0f, -1.0f, -2.0f, -2.0f, -3.0f, -3.0f, -4.0f, -4.0f, -5.0f, -5.0f, -6.0f, -6.0f, -7.0f, -7.0f, -8.0f, -8.0f));
 
 __m512 test_mm512_mask_movehdup_ps(__m512 __W, __mmask16 __U, __m512 __A) {
   // CHECK-LABEL: test_mm512_mask_movehdup_ps
@@ -8702,6 +8704,7 @@ __m512 test_mm512_moveldup_ps(__m512 __A) {
   // CHECK: shufflevector <16 x float> %{{.*}}, <16 x float> %{{.*}}, <16 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14>
   return _mm512_moveldup_ps(__A);
 }
+TEST_CONSTEXPR(match_m512(_mm512_moveldup_ps((__m512){+1.0f,-1.0f,+2.0f,-2.0f,+3.0f,-3.0f,+4.0f,-4.0f,+5.0f,-5.0f,+6.0f,-6.0f,+7.0f,-7.0f,+8.0f,-8.0f}), +1.0f, +1.0f, +2.0f, +2.0f, +3.0f, +3.0f, +4.0f, +4.0f, +5.0f, +5.0f, +6.0f, +6.0f, +7.0f, +7.0f, +8.0f, +8.0f));
 
 __m512 test_mm512_mask_moveldup_ps(__m512 __W, __mmask16 __U, __m512 __A) {
   // CHECK-LABEL: test_mm512_mask_moveldup_ps

@github-actions GitHub Actions
Copy link

github-actions bot commented Aug 6, 2025

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command: 
git-clang-format --diff HEAD~1 HEAD --extensions h,c -- clang/lib/Headers/avx512fintrin.h clang/lib/Headers/avxintrin.h clang/test/CodeGen/X86/avx-builtins.c clang/test/CodeGen/X86/avx512f-builtins.c
View the diff from clang-format here. 
diff --git a/clang/lib/Headers/avx512fintrin.h b/clang/lib/Headers/avx512fintrin.h
index b743c5b84..87f9d5f5a 100644
--- a/clang/lib/Headers/avx512fintrin.h
+++ b/clang/lib/Headers/avx512fintrin.h
@@ -5304,8 +5304,7 @@ _mm512_mask_store_epi64 (void *__P, __mmask8 __U, __m512i __A)
 }
 
 static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR
-_mm512_movedup_pd (__m512d __A)
-{
+_mm512_movedup_pd(__m512d __A) {
   return (__m512d)__builtin_shufflevector((__v8df)__A, (__v8df)__A,
                                           0, 0, 2, 2, 4, 4, 6, 6);
 }
@@ -8666,8 +8665,7 @@ _mm512_mask_testn_epi64_mask (__mmask8 __U, __m512i __A, __m512i __B)
 }
 
 static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR
-_mm512_movehdup_ps (__m512 __A)
-{
+_mm512_movehdup_ps(__m512 __A) {
   return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
                          1, 1, 3, 3, 5, 5, 7, 7, 9, 9, 11, 11, 13, 13, 15, 15);
 }
@@ -8689,8 +8687,7 @@ _mm512_maskz_movehdup_ps (__mmask16 __U, __m512 __A)
 }
 
 static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR
-_mm512_moveldup_ps (__m512 __A)
-{
+_mm512_moveldup_ps(__m512 __A) {
   return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
                          0, 0, 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12, 14, 14);
 }
diff --git a/clang/lib/Headers/avxintrin.h b/clang/lib/Headers/avxintrin.h
index 0ccdaaa1f..21196a2fd 100644
--- a/clang/lib/Headers/avxintrin.h
+++ b/clang/lib/Headers/avxintrin.h
@@ -2393,8 +2393,7 @@ _mm256_cvtss_f32(__m256 __a)
 /// \returns A 256-bit vector of [8 x float] containing the moved and duplicated
 ///    values.
 static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR
-_mm256_movehdup_ps(__m256 __a)
-{
+_mm256_movehdup_ps(__m256 __a) {
   return __builtin_shufflevector((__v8sf)__a, (__v8sf)__a, 1, 1, 3, 3, 5, 5, 7, 7);
 }
 
@@ -2418,8 +2417,7 @@ _mm256_movehdup_ps(__m256 __a)
 /// \returns A 256-bit vector of [8 x float] containing the moved and duplicated
 ///    values.
 static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR
-_mm256_moveldup_ps(__m256 __a)
-{
+_mm256_moveldup_ps(__m256 __a) {
   return __builtin_shufflevector((__v8sf)__a, (__v8sf)__a, 0, 0, 2, 2, 4, 4, 6, 6);
 }
 
@@ -2440,8 +2438,7 @@ _mm256_moveldup_ps(__m256 __a)
 /// \returns A 256-bit vector of [4 x double] containing the moved and
 ///    duplicated values.
 static __inline __m256d __DEFAULT_FN_ATTRS_CONSTEXPR
-_mm256_movedup_pd(__m256d __a)
-{
+_mm256_movedup_pd(__m256d __a) {
   return __builtin_shufflevector((__v4df)__a, (__v4df)__a, 0, 0, 2, 2);
 }

phoebewang
phoebewang approved these changes Aug 7, 2025
View reviewed changes
Copy link
Contributor

@phoebewang phoebewang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

@RKSimon RKSimon merged commit 6abf4f3 into llvm:main Aug 7, 2025
12 of 13 checks passed
@RKSimon RKSimon deleted the x86-movedup-constexpr branch August 7, 2025 07:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@phoebewang phoebewang phoebewang approved these changes

Assignees
No one assigned
Labels
backend:X86 clang:headers Headers provided by Clang, e.g. for intrinsics clang Clang issues not falling into any other category
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@RKSimon @llvmbot @phoebewang